A foreign brand that finally lands on a Chinese platform often treats customer reviews as a support afterthought. The listing goes live, orders start arriving, and the star rating is left to chance. That is a costly mistake. On Tmall Global, JD Worldwide, RED and Douyin, the review section is not a footnote to the sale; it is part of the sales engine itself. Shoppers read before they buy, the platform algorithm reads what they write, and both readings decide whether the next shopper ever sees the product at all.
This guide explains why review management belongs on the marketing side of the house rather than the support side, which metrics actually move sales, how a brand should respond when a review goes wrong, what practices are strictly off limits, and how to build a review routine that survives the realities of cross-border operations. It is written for operators who already sell or are about to sell into China, and who want a system they can actually keep running.
Why reviews are a traffic engine, not just feedback
Most Western brands think of reviews as a post-purchase courtesy: a place where happy customers say thanks and unhappy ones vent. On Chinese platforms the mechanism is different. Search ranking, category placement and paid-traffic efficiency are all influenced by the review profile of a listing. A product with a thin or stale review history is quietly pushed down, while a product that keeps earning fresh, detailed reviews is shown to more people for the same advertising spend.
The reason is mechanical. The platforms want to show results that convert, and review velocity is one of the strongest signals they have that a product is alive and trusted. Roughly 70% of Chinese shoppers read reviews before purchasing, and a listing that cannot show recent, substantive feedback looks abandoned even when it is not. A one-star drop on the average, or a long gap with no new reviews, can cut conversion by as much as 20–30% without any change to the product or the price.
That is why review management is a marketing function. It shapes who sees the product, not just how the people who already bought feel about it. Treating it as a customer-service ticket queue leaves the traffic on the table.
The metrics that actually matter
Not every review number is worth a dashboard. Four metrics explain most of the commercial effect, and a brand should watch them weekly rather than annually.
Review velocity is the number of new reviews a listing earns per week. A listing that adds 30–50 new reviews per week tends to hold or improve its search position, while one that stalls tends to drift down. Velocity matters more than total count, because platforms weight recent activity heavily.
Average star rating is the gate. On Tmall and JD, products sitting at 4.8 stars or above with review counts in the thousands consistently outrank comparable listings with thinner histories. Below about 4.5, paid traffic gets expensive and organic reach contracts.
Photo and video review rate measures how many reviews include an image or clip. These reviews carry far more persuasive weight than text alone, and platforms surface them preferentially. A brand should track the share of reviews that are visual, not just the count.
Response rate and response time measure how often and how fast the brand replies to reviews, especially negative ones. Platforms reward visible, timely merchant responses, and shoppers read the responses as much as the original complaints. A brand that answers within 48–72 hours signals that a real operator stands behind the product.
A response playbook for negative reviews
Negative reviews are inevitable, and pretending otherwise leads to worse outcomes. The goal is not to delete the complaint but to demonstrate, in public, that the brand handles problems competently. A simple four-step playbook works across platforms.
Acknowledge specifically and without defensiveness. A generic apology reads as scripted; a reply that references the actual problem shows a human read it. Investigate before promising a fix, because a wrong assurance damages credibility more than a slow one. Resolve through the channel the platform provides, whether that is a refund, a replacement or a direct message, and keep the public thread updated so the sequence is visible. Follow up after the resolution to confirm the customer is satisfied, which often prompts them to edit or append a note.
Timing is part of the message. A response that lands inside the 48–72 hour window reads as attentive; one that arrives after a week reads as indifference regardless of its content. Cross-border brands should pre-authorise a local responder or a clear escalation path so the clock does not expire while the headquarters sleeps.
What foreign brands must never do
Chinese platform rules and advertising law are strict about review integrity, and the penalties scale from listing takedown to store closure. Four practices are firmly off limits.
Buying fake reviews or hiring agencies that post them is the most dangerous. Detection is improving, and a fabricated-review scheme can erase months of legitimate progress in a single enforcement action. Incentivising silence, by offering refunds or gifts in exchange for deleting or not posting a negative review, is also prohibited and equally visible to enforcers. Fabricating ratings through shell accounts, coordinating review bursts with related parties, or copying reviews from other markets onto a China listing all fall into the same forbidden category.
The safer path is to earn reviews honestly and respond to the bad ones well. A 4.8-star listing built on real feedback outlasts a 4.9-star listing built on fabricated ones, because the second one can be removed overnight.
Building a review-acquisition routine
Reviews do not appear by wishing. A cross-border brand needs a routine that fits thin local staffing and time-zone gaps. Three moves cover most cases.
Ask at the right moment. The request should arrive after the product has been in the customer’s hands long enough to form an opinion but before the memory fades, typically a few days after confirmed delivery. A well-timed prompt on the platform’s own message system outperforms a generic blast.
Make the act effortless. Deep-link the review form, pre-fill what the platform allows, and thank the customer for the time. The easier the step, the higher the completion rate, and a higher completion rate is what lifts velocity toward the 30–50 per week range.
Reward the behaviour you want without paying for the verdict. Points, small perks or loyalty credit for leaving a review are acceptable in many contexts; conditioning the reward on a positive score is not. Keep the incentive attached to the act of reviewing, never to the star result.
How to measure the program
A review program is only worth running if it is measured, and measured against the metrics that move sales. A monthly scorecard should track velocity, average star, photo or video review share, response rate within the 48–72 hour window, and the share of negative reviews that were resolved and then updated by the customer.
The leading indicator is velocity, because it predicts ranking and reach. The lagging indicator is the average star and its trend, because that reflects whether the underlying product and fulfilment are sound. When velocity rises but the star average falls, the brand has a product or logistics problem, not a review problem, and the fix belongs upstream rather than in the review tool.
Review management is not glamorous, but on Chinese platforms it is one of the highest-leverage marketing activities a foreign brand can run. The brands that win are not the ones with the most reviews on day one; they are the ones that treat every review as a small piece of public, searchable, trust-building content, and that answer the bad ones before the algorithm and the next shopper decide for them.